Some data systems employ multiple target devices (e.g., storage devices) which may be accessed by initiators of host systems through one or more expanders in order to read and write data. The expanders interconnect the target devices and the initiators via physical interfaces, or “Phys”, to increase the number of target devices that the initiators may access. A protocol such as SAS uses a series of commands to communicate between devices in the system. Sometimes, a transmission of commands is interrupted in a way such that a receiving device is unable to interpret and respond to a transmitting device. This may result in a situation known as deadlock where the receiving device and transmitting device are no longer able to understand and communicate with one another. Path recovery is used within a SAS topology to avoid deadlock by deterministically choosing which partial pathway(s) to tear down to allow at least one connection to complete. However, the path recovery process begins after an expiration of a partial pathway timeout which inherently delays the process of deadlock detection and slows system performance. Furthermore, pathway recovery is sometimes initiated even when true deadlock conditions do not exist.